connectivity

Document created by cdnadmin on Jan 25, 2014
Version 1Show Document
  • View in full screen mode
This document was generated from CDN thread

Created by: Raymond Armstrong on 15-05-2012 08:49:39 AM
1. Have been trying to set up server via PCIe but cannot get the internal link to come up following procedure
interface ucse2/0
 ip address 69.184.252.1 255.255.255.240
 imc ip address 192.168.1.91 255.255.255.0 default-gateway 192.168.1.60
 imc access-port dedicated
RedHat rel 6 on server
Server eth0 set as follows:
DEVICE="eth0"
HWADDR="F0:F7:55:12:01:BE"
ONBOOT="yes"
BOOTPROTO="static"
IPADDR="69.184.252.3"
#ETHTOOL_OPTS="speed 1000 duplex full space autoneg on" <tried with this line in and out of the config>
NETMASK="255.255.255.240"
We are seeing no link between the router and the server on the PCIe connection - is there anything we are missing?

2. Also tried using MGF Vlan following procedure
interface ucse2/1
 description Internal switch interface connected to Service Module
 switchport trunk allowed vlan 1,2,100,1002-1005
 switchport mode trunk
 no ip address
!
interface Vlan1
 no ip address
!
interface Vlan100
 ip address 69.184.252.17 255.255.255.240
!
DEVICE="eth0"
HWADDR="F0:F7:55:12:01:BE"
ONBOOT="yes"
BOOTPROTO="static"
IPADDR="69.184.252.19"
#ETHTOOL_OPTS="speed 1000 duplex full space autoneg on" <tried with this line in and out of the config>
NETMASK="255.255.255.240"
r107476#show ip int br
Interface                  IP-Address      OK? Method Status                Protocol
Embedded-Service-Engine0/0 unassigned      YES NVRAM  up                    up     
GigabitEthernet0/0         172.16.1.2      YES NVRAM  down                  down   
GigabitEthernet0/1         69.184.252.33   YES NVRAM  up                    up     
GigabitEthernet0/2         135.0.13.34     YES NVRAM  up                    up     
ucse2/0                    69.184.252.1    YES NVRAM  administratively down down   
ucse2/1                    unassigned      YES unset  up                    up     
Loopback0                  135.2.5.181     YES NVRAM  up                    up     
NVI0                       135.2.5.181     YES unset  up                    up     
Vlan1                      unassigned      YES unset  up                    up     
Vlan100                    69.184.252.17   YES manual down                  down 

Unsure how the vlan can come up without an interface placed in vlan 100? again we are seeing no link

3. We tried configuring the server with a direct cable between GE2 and the router GE0/1 and this would ping intermittently and neither the router or the server could see the arp of the other device.
r107476#ping 69.184.252.35
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 69.184.252.35, timeout is 2 seconds:
*May 15 11:35:34.812: IP ARP: sent req src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.35 0000.0000.0000 GigabitEthernet0/1.
*May 15 11:35:36.812: IP ARP: sent req src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.35 0000.0000.0000 GigabitEthernet0/1.
*May 15 11:35:38.812: IP ARP: sent req src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.35 0000.0000.0000 GigabitEthernet0/1.
*May 15 11:35:40.812: IP ARP: sent req src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.35 0000.0000.0000 GigabitEthernet0/1.
*May 15 11:35:42.812: IP ARP: sent req src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.35 0000.0000.0000 GigabitEthernet0/1.
Success rate is 0 percent (0/5)
r107476#
We then swopped to GE3 > GE0/1 on the router and we could ping and ssh before everything locked up.
r107476#ping 69.184.252.35
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 69.184.252.35, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/1 ms
r107476#ping 69.184.252.35
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 69.184.252.35, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/1 ms
r107476#ssh -l root 69.184.252.35
Password:
Last login: Thu Mar 24 12:23:46 2005 from 69.184.252.33
[root@cisco3925 ~]# ls
anaconda-ks.cfg  Documents  install.log         Music     Public     Videos
Desktop          Downloads  install.log.syslog  Pictures  Templates
[root@cisco3925 ~]# ls /
bin   cgroup  etc   lib    lost+found  misc  net  proc  sbin     srv  tmp  var
boot  dev     home  lib64  media       mnt   opt  root  selinux  sys  usr
[root@cisco3925 ~]#
***************locked up*************
We also noted the following:
r107476#
*May 15 12:07:53.303: old: SERVICE_MODULE_STATE_ERRQ new: SERVICE_MODULE_STATE_SHDN
r107476#

r107476#
*May 15 12:13:58.875: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/1, changed state to down
*May 15 12:13:59.875: %LINK-3-UPDOWN: Interface GigabitEthernet0/1, changed state to down
r107476#
*May 15 12:14:01.615: IP ARP rep filtered src 69.184.252.35 f0f7.5512.01c1, dst 69.184.252.35 f0f7.5512.01c1 wrong cable, interface GigabitEthernet0/1
*May 15 12:14:02.875: %LINK-3-UPDOWN: Interface GigabitEthernet0/1, changed state to up
r107476#
*May 15 12:14:02.875: IP ARP: sent rep src 69.184.252.33 70ca.9ba3.0c81,
                 dst 69.184.252.33 ffff.ffff.ffff GigabitEthernet0/1
*May 15 12:14:03.631: IP ARP req filtered src 69.184.252.35 f0f7.5512.01c1, dst 69.184.252.35 ffff.ffff.ffff wrong cable, interface GigabitEthernet0/1
*May 15 12:14:03.875: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/1, changed state to up
r107476#
r107476#ping 69.184.252.35
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 69.184.252.35, timeout is 2 seconds:
.!!!!
Success rate is 80 percent (4/5), round-trip min/avg/max = 1/1/1 ms
r107476#
r107476#ssh -l root 69.184.252.35
Password:
Last login: Thu Mar 24 12:33:40 2005 from 69.184.252.33
[root@cisco3925 ~]# more /etc/sysc
r107476(config-if)#switchport
*May 15 12:17:53.303: old: SERVICE_MODULE_STATE_SHDN new: SERVICE_MODULE_STATE_FAIL
 
we are seeing NO BUFFER SPACE AVAILABLE on the server

we are currently looking at the drivers and configurations for point 3.

Subject: RE: connectivity
Replied by: Raymond Armstrong on 15-05-2012 11:44:41 AM
We belive the issue with 3. is due to the broadcom driver on the external gig ports - we have tried updating the kernal and the drivers but still cannot get this working.

Subject: RE: connectivity
Replied by: Brett Tiller on 15-05-2012 02:18:22 PM
Hi Raymond,

Regarding your configuration in the UCSE 2/0 interface , "access-port dedicated" means that you have an ethernet cable connection to the 'M' port on the ucse module which would then connect to your switch.  This connection separates packets going to the host bypassing the router.   If you desire to use the PCIE connection then you would remove that Ethernet connection and connect to the GE2 port on the front panel and connect the other end of that cable to the switch.  The access-port command should then be changed to "access-port host GE2".  After you've changed this configuration, please power down the router completely and reload it.   This reboot may clear up the status changes that you are experiencing.

After you've completed the steps above, if you are unable to connect to the RedHat OS, please add an ip route in the router specifying the ip of this server linking it to the module interface.  I've provided an example below.

ip route 69.184.252.3  255.255.255.255 ucse 2/0

If you continue to see problems connecting to Red Hat please check that LInux SE is disabled, and that there is no Red Hat firewall blocking the connection.  Failing that, please attach your router configuration and let us know the Ethernet connections as well.

Thanks,

Brett

Subject: RE: connectivity
Replied by: Raymond Armstrong on 16-05-2012 04:54:56 AM
Hi Brett - current status

Test 1 PCIe
what I am actually trying to do for is use the PCIe layer3 connection have removed

imc ip address 192.168.1.91 255.255.255.0 default-gateway 192.168.1.60
imc access-port dedicated

So if we strip the above out are we saying that the PCIe connection for GE0 should act like a normal "GE" connection?


Test 2 MGF vlan
further investigation required

Test 3 using external cabling GE2/3
We may have got this working and can now connect using the latest kernel though we need to understand whether the kernel is using the same broadcom drivers as the version on rel 6.0 or whether it is now using the bnx2 broadcom drivers - further testing and analysis required.

Subject: RE: connectivity
Replied by: Raymond Armstrong on 16-05-2012 05:56:38 AM
Test 1 PCIe
with the following removed
imc ip address 192.168.1.91 255.255.255.0 default-gateway 192.168.1.60
imc access-port dedicated
it now appears to work

Test 2 MGF vlan
changed
switchport mode to access / access vlan 100 and this now works

Test 3 using external cabling GE2/3
further testing and analysis required.

Subject: RE: connectivity
Replied by: Manuka Jayakody on 16-05-2012 01:34:29 PM
Hi Raymond,

Redhat's eth0 (PCIe) connectivity should work regardless of the imc ip address and imc access-port commands. you may need to add a route to the eth0 ip address as Brett pointed out.

I've been using Redhat's eth1 (MGF) connectivity with vlan1. for eg:

interface Vlan1
ip address 100.1.1.1 255.255.255.0
---
ip route 100.1.1.2 255.255.255.255 Vlan1

--
100.1.1.2 is the eth1 address and using eth1 default-gateway as 100.1.1.1


Redhat has just posted the latest signed driver that works with ucse, per our request. If you still see driver issues you can download it from here and install it

https://rhn.redhat.com/network/software/packages/name_overview.pxt?package_name=kmod-tg3&archIdList=520&archLabelList=x86_64&search_subscribed_channels=

Attachments

    Outcomes